The aim of this study was to assess agreement between observers in performing anthropometric measures and a subjective assessment of nutritional status in hospitalized adults. In a general university hospital, independent observers obtained patients' current weight, usual weight, height, arm circumference, triceps skin fold thickness, percentage of weight loss, Body Mass Index, muscle arm circumference, and Subjective Global Assessment. Agreement was analyzed using the Kappa coefficient, paired t test, and intra-class correlation coefficient. The sample consisted of 102 patients, whose mean age was 53±15 years. Good correlations (ICC>0.86) and low mean differences between observers were obtained for all anthropometric measures. Agreement in classifying patients into categories of Body Mass Index was very strong (kappa=0.98; 95%CI:0.95-1.0), weak for Subjective Global Assessment (kappa=0.46; 95%CI:0.31-0.60) and moderate for the 5th and 10th percentiles of arm circumference (kappa=0.78; 95%CI:0.65-0.91), muscle arm circumference (kappa=0.73; 95%CI:0.54-0.92) and triceps skinfold thickness (kappa=0.65; 95%CI:0.48-0.82). Several patients had clinically relevant discrepancies in percentage of weight loss, arm circumference, triceps skin-fold thickness, and muscle arm circumference. Although anthropometric methods showed good statistical reproducibility, we should be prudent in applying these measures in the hospital setting, since discrepancies were observed, especially in arm anthropometry and Subjective Global Assessment.
Reliability; Reproducibility; Anthropometry; Nutrition assessment; Nutritional status
